Research

Understanding and designing thermal materials from molecular motion to data-driven discovery.

My research connects molecular simulation, thermal transport theory, machine learning, and emerging experimental workflows to explain how materials store, move, and dissipate heat.

Current Focus at MIT

Heat-conducting liquids as a platform for thermal energy science.

Liquids are central to cooling, thermal storage, batteries, electronics, and energy systems, yet they remain harder to model than gases or crystals because molecular vibration, diffusion, rotation, and configurational rearrangement all contribute at once. My current work asks how these molecular motions determine liquid heat capacity and thermal conductivity.

In the NanoEngineering Group at MIT, I am studying heat-conducting liquids through theory, molecular dynamics, lattice-dynamics-inspired analysis, and experimental concepts for fast thermal-property screening. A key direction is separating vibrational energy storage from configurational contributions such as molecular rearrangement and hydrogen-bond exchange.

Related Projects Publications

Research Program

Four connected directions

Across these projects, I use computation to connect molecular mechanisms with engineering-scale material choices.

01 | Liquid thermal physics

Heat capacity and thermal conductivity of liquids

I develop models and simulation analyses for liquid thermal properties, with emphasis on how vibration, rotation, diffusion, and configurational changes combine to determine heat storage and transport.

  • Heat-conducting liquids and thermal energy systems.
  • Instantaneous normal modes and local molecular motion projections.
  • Frequency-domain thermoreflectance concepts for liquid-property measurement.

02 | Polymer informatics

AI-driven polymer property prediction

I build workflows that combine molecular simulation, group contribution estimates, experimental data, and graph-based learning to predict polymer properties and screen large chemical spaces.

  • ADEPT for automated polymer simulation data generation.
  • PolyGraphMT for multi-task, multi-fidelity polymer learning.
  • Transfer learning for polymer specific heat capacity.

03 | Nanoscale transport

Thermal resistance at solid-liquid interfaces

My earlier work quantified Kapitza resistance and interfacial thermal transport in water-graphene, water-CNT, and hBN-water systems, revealing how confinement, electrostatics, hydrogen bonding, and surface structure control heat flow.

  • Equilibrium molecular dynamics for interfacial thermal resistance.
  • Nanoconfinement effects in carbon nanotubes.
  • Electrostatic control of hBN-water thermal interfaces.

04 | Transport tools and devices

Simulation tools for nanofluidic and energy materials

I develop and apply computational tools for constrained molecular systems, nanochannel slip and interfacial transport, AC electrohydrodynamic flow, and coupled thermal-electrical behavior in energy-storage materials.

  • NAVIS for velocity slip, thermal slip, and interfacial transport.
  • Dofulator for degrees of freedom in constrained molecular systems.
  • Molecular insight into EHD flow and supercapacitor electrodes.

Approach

Physics-grounded computation, not black-box prediction.

Molecular mechanisms

I use atomistic simulation to identify the motions, interactions, and interfacial structures responsible for observed transport behavior.

Predictive models

I translate simulation and experimental signals into machine-learning models that remain tied to physical descriptors and uncertainty-aware workflows.

Scalable platforms

I build reusable tools and datasets so individual simulations can become general workflows for discovery, screening, and validation.

Methods

Computational and experimental toolset

LAMMPS VMD Packmol MDAnalysis Pysimm Psi4 Python PyTorch TensorFlow RDKit Optuna Graph Neural Networks Transfer Learning Ansys Fluent Frequency-Domain Thermoreflectance High-Throughput Screening